Can the Third Part of Duodenum Behind SMA Be Detected With Confidence on CT as a Proposed Mechanism for Imaging Suspected Malrotation in Children? A Preliminary Proof of Concept Study

Abstract

The aims of this study were to determine frequency and reliability of computed tomography (CT) detection of anatomic landmarks for imaging suspected midgut malrotation in infants and children, and to calculate an estimated effective dose of an upper abdominal CT scan in our patient population. Fifty consecutive pediatric patients who underwent a CT scan that included their upper abdomen between August 2016 and February 2018 were included. Four pediatric radiology consultants independently reviewed CT scans for detection of the third part of the duodenum and defined their confidence level of this through identification of continuity with the pyloric antrum, D1, D2, and D4 components of the duodenum, as well as the duodenojejunal flexure.Interobserver variability was assessed using Fleiss κ for agreement. A dose estimate, per scan, was calculated using the scanner dose-length product and published conversion factors by Deak. Thirty patients were boys. The average age was 7.5 ± 5.4 years (6 days to 16 years). The D3 segment was definitely identified in 70% of scans, with 68% to 73%, moderate agreement between the readers and a Fleiss κ of 0.47 to 0.52. The DJ flexure was definitely identified in only 30.5% cases, with 35%, poor agreement between readers (Fleiss κ of 0.03). The average estimated dose for a targeted CT scan of the abdomen was 0.9 mSv (0.04-2.4 mSv). The third part of the duodenum, which is integral in excluding malrotation on cross-sectional studies, was "definitely" identified in 70% of CT scans of children in our study, with 68% to 73% agreement between the readers and a Fleiss κ of 0.47 to 0.52.These preliminary proof of concept results demonstrating a combination of a comparable CT dose in relation to upper gastrointestinal contrast studies and an acceptable number of cases delineating the third part of the duodenum with moderate agreement are a first step in suggesting low-dose CT for an imaging diagnosis of malrotation. Malrotation can be excluded in cases where D3 is well demonstrated in the normal position, which negates the need to automatically refer children with bilious emesis to specialist centers for upper gastrointestinal contrast studies.